human/nlp/nlp_split.py

#encoding = utf8
import re
from abc import ABC, abstractmethod


class NLPSplit(ABC):
    @abstractmethod
    def handle(self, message: str):
        pass


class PunctuationSplit(NLPSplit):
    def __init__(self):
        self._pattern = r'(?<!\d)[,.，。？！：；、]'

    def handle(self, message: str):
        message = message.replace('*', '')
        message = message.replace('#', '')
        match = re.search(self._pattern, message)
        if match:
            pos = match.start() + 1
            msg = message[:pos]
            msg = msg.strip()
            message = message[pos:]
            return message, msg
        return message, ''